Step 4: data gathering and analysis

7
Potential causes and data collection and analysis
The analysis of data is another very important step since the main goal is to gather
previous sources of information for the sake of reviewing if the failure has previously
occurred and, of course, to understand if the corrective actions were ineffective since
the event may be related to past failures and possible ineffective action. Therefore, the
goal of data analysis and data modeling is to identify useful information, which could be
used for suggesting conclusions and, more importantly, to supporting decision-making
based on factual data sources. This can only be achieved if the organization establishes
procedures for collecting the right type of data, for analyzing data, and by identifying up
front the techniques for interpreting the results of such procedures.
Data analysis has multiple approaches, and many techniques can be used. Of course,
to do this, the organization needs to identify what data is critical to their business or
processes, and to assign responsibilities for the ongoing data collection and data anal-
ysis. This could be as simple as data collection to maintain histograms over specific
periods. By implementing an effective data collection program, whenever data needs
to be reviewed during an investigation, the process will be painless and more efficient
without the need for additional resources or time.
Data analysis should be planned to include appropriate statistical methodology and
should be employed where necessary to detect and address recurring quality problems.

Data, quality indicators, and triggers

The need for correction and corrective actions will be determined when a quality indi-
cator has triggered, where a risk assessment determines that action must be taken,
or while conducting a review of each quality indicator’s trend during the corrective
and preventive action (CAPA) review meeting. In addition, management should also
initiate CAPA requests for items they determine significant. The ultimate decision for
initiating a CAPA should rest with the senior quality manager.

Levels of reviews
• At the business level, key indicators should be used to determine whether quality objectives
are being achieved, and internal audits assess the suitability and effectiveness of the quality
management system.
• At the process level, data is analyzed from a specific product, and process-monitoring sources
are typically defined in CAPA procedures and/or production and process control procedures.

CAPA in the Pharmaceutical and Biotech Industries. http://dx.doi.org/10.1016/B978-1-907568-58-9.00007-2

Copyright © 2016 by J. Rodriguez. Published by Elsevier Ltd. All rights reserved.
74 CAPA in the Pharmaceutical and Biotech Industries

• At the workstation level, routine monitoring of key parameters should be planned and imple-
mented as required by an approved control plan, quality plan, or protocol.

Analysis of source quality data will include global analysis across all source CAPA
data lines (i.e., analysis of common factors, such as problem, product, lot number,
staff, and other criteria against all corresponding source CAPA data, including qual-
ity complaints, nonconformances, internal and external audits, and CAPA reports) to
identify any additional existing and potential quality problems.

Data collection should be planned as follows

Data owners may report independently within their department or through interface
with the Quality Assurance group. The Quality Assurance/Regulatory department is
usually responsible for publishing CAPA trigger matrix meeting reports in support of
the Quality Scorecard. This effort typically involves direct interface with data own-
ers who may provide data collection requirements and who may provide comments
needed for management review of adverse trends. Meetings should be held monthly or
quarterly to review quality indicator results with emphasis on areas that fail to achieve
targets or that have unfavorable trends.

Data analysis and the investigation process

At the beginning of the investigation, a list of possible causes is created. This will form
the basis for collecting relevant information, test data, and provide the investigation
owner with information to decide the most appropriate investigation methods to be
utilized. For example, consider the situation where a large batch of parts from a mon-
itor final testing phase was discovered to be nonconforming. There are many possible
causes for this condition including operator error, incorrect software, an uncalibrated
tool, a material problem, or a design problem. By considering possible causes, appro-
priate information and data can be collected that will ultimately be used to determine
the origin of the problem.

Data collection
A data collection program should be established up front, and this program should
describe the basis for collecting relevant information, assigned responsibilities for the
collection, and ongoing data review and analysis. There are many possible causes for
ongoing failure situations including equipment wear and tear, operator error, incor-
rect software, a dull or broken tool, incorrect or obsolete procedures and drawings,
material problems, design problems, off-label use of the product, etc. By considering
all possible causes, appropriate information and data can be collected that will be ulti-
mately used to determine the root cause of the problem.
Step 4: data gathering and analysis 75

Potential causes and data collection and analysis

Typically at the beginning of the investigation, a list of possible causes is created. This
will form the basis for collecting relevant information, testing data, and providing the
investigation owner with enough information that could be used to determine recur-
rence rates.
For example, during a complaint investigation a review of recurrence of the failure
should include a time frame. How would you determine the appropriate time frame?
You can start by looking at the manufactured dates for the products that had the same
or similar type of failures. This will provide a time period.

Results and data

The results of the data collection must be documented and organized. This may include
a combination of testing results and/or a review of records, processes, service informa-
tion, design controls, operations, and any other data that may lead to a determination
of the fundamental cause of the problem.
The resulting documentation should be complete and address all of the possible
causes that were previously determined. This information is used to determine the root
cause of the problem.
Begin by looking at the methods to be used to collect the information. For the
CAPA program to work, the data to be analyzed must be appropriate to the system.
Ensure that the data is sound, that it is not junk data—and, last year’s statistics have
limited value for the company. Look at data on a timely basis. Identify who will look
at the data, determine how you are going to measure it, and monitor the activities that
are going to be used for a CAPA prospective data analysis. Determine when you will
perform the analysis or the evaluation of the data collection activities.
One important element within the CAPA program is the collection and analysis of
data to identify all the things that could be a potentially nonconforming product issue.
For example, let us say we have noticed reworking of procedures or reworking of
product. Those are telling quality indicators that we should not ignore. The existence
of reworking indicates a quality problem; if you rework, you have a problem. Either
you are not manufacturing the way you are supposed to or you are deviating from the
process. Under these circumstances, you want to make sure that within the CAPA
program you require an investigation.

Gather information and analyze data

Once the problem has been identified, then we start gathering data for the data analy-
sis. This is the beginning of the investigation for the root cause. Data analysis begins
by considering what information is needed. We do the analysis and gather the data as
interrelated steps. We may go back and forth between the two steps, especially at the
beginning, while we are discovering what information is needed. Go back to the well
of information until you have everything you feel answers the outstanding questions.
76 CAPA in the Pharmaceutical and Biotech Industries

You may be looking at batch records, DHRs, complaints, or service reports.

You are gathering the information you need to perform the data analysis, which
is the third step in the program. As you are doing the investigation, you must ask
the following: what are the procedures or what are the processes involved, what is
the equipment, what has changed, and are these correct procedures? Keep asking
questions until you have satisfied every open issue to obtain the information you
need and have a clear understanding of the event; you may start interrogating
people—talking to people who work on the processes or procedures. Keep asking
questions to get all the information and documentation you feel is necessary to
perform an effective root cause analysis.
Once you have all the data in front of you, even as you are analyzing the data,
things may be falling into place. You should be thinking of different ways to look
at the data. Try a fault tree analysis, or a fishbone, or other chart types to look at
the data in new ways. Each of these tools can help you see the information in a
different light.

CAPA system data collection and analysis

The purpose of this section is to outline the requirements for collecting and analyzing
data from the CAPA system to achieve consistency in the approach, evaluation, anal-
ysis, and reporting of quality data.

The general process for data collection and analysis

Step 1
• Identify proper quality data owners review metrics for CAPA data review meetings and
provide an explanation of the data utilizing the appropriate method of analysis. This must be
defined in the company’s data trends procedure or a corrective and preventive action system
data and analysis procedure
• This analysis should include, at minimum, the following:
• Identification of any known or suspected causes
• Any additional issues or concerns

Step 2
The data owner should compile the completed metrics in preparation for the CAPA
system review meetings.

Step 3
The quality data owners, at the CAPA system review meeting, present quality metrics
and analysis.
Step 4: data gathering and analysis 77

Step 4
The CAPA Review Board reviews the quality metrics and analysis to determine the
need for further action, if required.
Note: Based on further analysis or new information, management should determine
that some metrics are no longer useful and replace them with other metrics or alter-
nate methods of analysis. The normal change order process should be followed when
changes to this document are made.

CAPA system review

To achieve consistency in the approach, evaluation, analysis, and reporting of quality
data, the following is essential and should be documented as appropriate:
• Analysis of the CAPA system is conducted monthly through the CAPA system review meet-
ings. The meetings should be conducted as follows.
• The CAPA coordinator should establish in advance and communicate to all participants the
exact dates, time, duration, and location.
• The CAPA coordinator should establish an agenda and prepare and maintain minutes.
• The CAPA coordinator should lead the meeting per the agenda, facilitate discussion, and
circulate the attendance sheet to attendees.
• The CAPA coordinator should confirm that the required attendees, or their designees, from
each functional area are present.
• Department heads should appoint designees in the event the department head is not able to
attend the meeting. The department head should document these designees.
• No more than two designees are allowed to be in attendance at the CAPA review meeting.
• Acceptable written documentation may include memos, e-mails, etc.
• Presenters from quality data sources and departments present processed data and analysis
for review.
• Presenters advise the CAPA System Review Board when additional resources are necessary
to complete corrective actions and/or preventive actions.

The CAPA System Review Board should review the following:

• Open action items from the previous meeting to analyze their status and timeliness. This
ensures that meetings are closing the loop on assignments made from the previous meetings.
• Review metrics to ensure that investigations and CAPAs are effective and are processed
within the timelines established.
• Review metrics for appropriateness and determine if additional metrics or the reporting fre-
quencies of metrics are adequate, if applicable.
• Discuss late/unresolved investigations and CAPAs and decide on further action.
• Identify if new investigations or CAPAs are required and assign them to individuals or func-
tional department(s) with due dates for completion.
• Examples of where additional action is indicated include, but are not limited to, the following:
• Perceived high risk or indication of high risk is identified through the use of risk manage-
ment tools.
• Data presented demonstrates an adverse trend.
78 CAPA in the Pharmaceutical and Biotech Industries

• These actions are to be documented in the CAPA system review meeting minutes.
• Determine the appropriateness of the CAPA review meeting frequency.

The CAPA coordinator should document any new action items related to corrective
and preventive actions and enter into the CAPA system.
The CAPA coordinator should publish and retain minutes of the meeting. Minutes
of the meeting are documented. The minutes should be approved by the CAPA Review
Board members that attended the CAPA system review meeting, and they are verified
by the CAPA coordinator.
The target date for publication of meeting minutes is 10 business days from the date
of the CAPA system review meeting.
The CAPA coordinator should prepare a summary report(s) for management review
meetings. This report(s) should include a summary of the statistical reports required
by the CAPA system data and analysis procedure.

Problem analysis
The problem analysis is the conclusive interpretation of the data gathered in the
previous steps. The analysis tools may vary from general problem-solving techniques
to very specific elaboration programs for the determination of the reason behind a
problem occurrence.

Analysis of data
Data should be collected and analyzed to demonstrate the suitability and effectiveness
of the quality management system and to evaluate where improvements can be made.
Data sources include but are not limited to the following:
• Feedback
• Nonconformance data
• Supplier data
• CAPA trends
• Test results
• Management review process
• Internal auditing

Improvement
The quality improvement process should be established based on data trends. This will
help a company to achieve the organization’s continuous improvement of culture and
values and improvement of setting goals and objectives.
Step 4: data gathering and analysis 79

Investigation process metrics

• Metrics collected from the investigation process are identified in the CAPA System Data and
Analysis Procedure TP0398.
• The CAPA coordinator (or designee) is responsible to produce reports that contain metrics
and analysis that determine if further action is required as defined in the company’s Correc-
tive and Preventive Action System Data and Analysis process.

Closed loop corrective action

• Investigation data is analyzed and forwarded to the CAPA system review meetings. The
results of this analysis and any further decision to take action are recorded in the meeting
minutes from the CAPA system review meetings.
• Items requiring further action are elevated to the management review.